A green innovation performance evaluation method and system based on dynamic weight optimization

Abstract

The application provides a green innovation performance evaluation method and system based on dynamic weight optimization, and belongs to the technical field of green innovation performance evaluation. The application solves the problems of few evaluation data sources, unclear weighting method and lack of correction mechanism of evaluation results in the prior art. The method comprises the following steps: multi-source data access, data standardization and semantic mapping, index construction and core index determination, index weight solving, data source weight solving, index fusion score calculation, green innovation performance comprehensive score calculation, evaluation result output and short board diagnosis, and feedback correction and parameter updating. The system comprises the following modules: a data access module, a standardization processing module, an index construction module, a weight solving module, a data reliability module, a comprehensive evaluation module, a result output module, a feedback correction module and a main control processor.

Description

Technical Field

[0001] This invention relates to a method and system for evaluating green innovation performance based on dynamic weight optimization, belonging to the field of green innovation performance evaluation technology. Background Technology

[0002] With the simultaneous advancement of green and digital transformation, corporate green innovation activities are characterized by diverse sources of input, complex influencing factors, and lagging results. Traditional green innovation performance evaluation methods mostly rely on financial statements, a limited number of output indicators, or single-round expert scoring. While existing methods were initially feasible, their practicality has become significantly insufficient in the current multi-source data environment.

[0003] On the one hand, existing solutions generally fail to cover the entire process of green innovation activities. Green innovation is not only reflected in the number of patents, revenue from green products, or emission reduction results, but also more profoundly in technological support capabilities, organizational investment capabilities, environmental constraints, and market and social responsiveness. If only a single indicator is used for measurement, the evaluation results are prone to bias and cannot accurately reflect the true level of a company's green innovation.

[0004] On the other hand, existing evaluation schemes still tend to be static in their weighting methods. Some schemes rely entirely on expert experience, which, while convenient to implement, leads to significant differences in judgment among experts and poor stability of results. Other schemes rely entirely on objective data distribution, which, while reducing subjective intervention, easily overlooks industry experience, policy context, and business priorities. More importantly, many existing technologies fail to differentiate the credibility of different data sources, often simply juxtaposing internal system data, external policy texts, and social media feedback. Because these data vary significantly in completeness, timeliness, traceability, and consistency, evaluation results often deviate from reality.

[0005] Furthermore, the dynamic adjustment mechanisms in existing solutions typically remain at the level of principle description, lacking a clear and easily implemented calculation process. Existing solutions only dynamically adjust the weights based on the domain to which the current object belongs, without clearly explaining how the weights are formed, normalized, or corrected based on subsequent business feedback, making it difficult to deploy the system stably in engineering. Summary of the Invention

[0006] To address the problems of limited evaluation data sources, unclear weighting methods, and lack of correction mechanisms for evaluation results in existing technologies, this invention proposes a green innovation performance evaluation method and system based on dynamic weight optimization.

[0007] The technical solution adopted by this invention to solve the above problems is as follows: This invention proposes a green innovation performance evaluation method based on dynamic weight optimization, comprising the following steps: Step 1: Receive multi-source data related to green innovation activities based on the main control processor; Step 2: Based on the multi-source data, perform data preprocessing using the main control processor to form a unified data matrix; Step 3: Based on the data matrix and the main control processor, determine the technical indicators, organizational indicators, environmental indicators and output indicators, and construct a green innovation performance evaluation indicator system. The indicator system includes technical indicators, organizational indicators, environmental indicators and output indicators, and determine the core indicator set from it. Step 4: Based on the core indicator set, generate subjective and objective weights for each evaluation indicator using the main control processor, and then weight and fuse the subjective and objective weights to obtain the comprehensive weight of each evaluation indicator. Step 5: Based on the integrity, timeliness, traceability and consistency of each data source, calculate the integrity score, timeliness score, traceability score and consistency score respectively based on the main control processor, and perform weighted summation according to the preset sub-item coefficients to obtain the original credibility value of each data source. Normalize the original credibility value to obtain the data source weight of each data source. Step 6: Based on the standardized values ​​of each data source on each evaluation indicator in the core indicator set and the data source weights, calculate the fusion score of each evaluation indicator based on the main control processor; Step 7: Calculate the comprehensive score of green innovation performance based on the main control processor, according to the comprehensive weight of each evaluation indicator, the fusion score of each evaluation indicator, the environmental adjustment item, and the comprehensive penalty item. Step 8: Based on the comprehensive score of green innovation performance, output the green innovation performance evaluation results based on the main control processor, and generate sub-indicator analysis results, main shortcomings indicators, explanation of reliable data sources, and improvement suggestions; Step 9: Based on subsequent actual business feedback, adjust the subjective and objective fusion coefficient, the credibility sub-coefficient of the data source, the adjustment coefficient of the environmental adjustment item, and the penalty coefficient of the comprehensive penalty item based on the main control processor.

[0008] Furthermore, the multi-source data in step 1 includes enterprise business data, R&D data, financial data, environmental governance data, policy text data, social feedback data, and image or log data; Step 1 specifically includes: Based on the main control processor, it receives operational data from enterprise business systems, R&D activity data from R&D systems, input and cost data from financial systems, energy consumption and emission reduction data from environmental governance systems, policy text data from policy databases, feedback data from social platforms, and extracts structured or semi-structured data from park terminals, image acquisition equipment, or log systems. Based on the source, type, collection time, and associated business objects of the multi-source data, the main control processor classifies and identifies the multi-source data to form an original dataset with source traceability information.

[0009] Furthermore, step 2 specifically includes: Step 2.1: Based on the missing values, outliers, and dimensional differences of the numerical data in the original dataset, perform missing value imputation, outlier correction, and dimensional normalization using the main control processor; Step 2.2: Based on the semantic content of the text data in the original dataset, perform semantic processing, keyword extraction, and rule mapping using the main control processor; Step 2.3: Based on the event records of the log-type data in the original dataset, perform event aggregation, frequency statistics, and time window alignment using the main control processor; Step 2.4: Based on the visual information of the image data in the original dataset, perform target feature extraction and structured representation using the main control processor; Step 2.5: Based on the processing results of Steps 2.1-2.4, the main control processor maps all types of data into standardized values ​​to obtain a data matrix.

[0010] Furthermore, step 3 includes: Based on the data items in the data matrix that reflect the enterprise's green innovation technology foundation, technical indicators are determined based on the main control processor. These technical indicators are used to characterize the digital foundation, equipment foundation, and technological synergy capabilities upon which the enterprise's green innovation depends. Based on the data items in the data matrix that reflect the company's green innovation investment, organizational indicators are determined based on the main control processor. These organizational indicators are used to characterize the company's human resource investment, financial investment, institutional investment, and management coordination capabilities in green innovation. Based on the data items in the data matrix that reflect the external constraints and feedback of the enterprise, environmental indicators are determined based on the main control processor. These environmental indicators are used to characterize policy drivers, social feedback, environmental ethics, and external governance constraints. Based on the data items in the data matrix that reflect the results of enterprise green innovation, output indicators are determined based on the main control processor. The output indicators are used to characterize green patents, green products, green services, resource utilization improvements and emission reduction results. Based on technical, organizational, environmental, and output indicators, a core indicator set is formed by selecting from the candidate indicator set using the main control processor. The core indicator set includes one or more of the following: number of new energy facilities, number of green patents granted, coverage of smart scenarios, proportion of R&D personnel, R&D investment intensity, maturity of corporate green governance, intensity of social green feedback, and strength of green policy support.

[0011] Furthermore, step 4 specifically includes: Based on the evaluation objectives and industry experience, subjective weights for each evaluation indicator are generated using one of the following methods: expert scoring, analytic hierarchy process, or preset business rules, based on the main control processor. The subjective weights reflect the experts' judgments on the relative importance of each indicator. Based on the sample data corresponding to the core indicator set, the objective weight of each evaluation indicator is calculated by the main control processor based on one of the following: indicator dispersion, information entropy, or variance contribution. The objective weight reflects the information difference and distinguishing ability of each indicator in the sample data. Based on the preset subjective and objective fusion coefficients, the subjective and objective weights are weighted and fused using the main control processor to generate the comprehensive weights for each evaluation indicator. w u ; u The overall weight of each indicator is: (1); In formula (1), w u Indicates the first u The final weight of each indicator, θ This represents the fusion coefficient between subjective and objective weights, with a value ranging from 0 to 1. For the first u The subjective weight of each indicator, For the first u The objective weight of each indicator.

[0012] Furthermore, step 5 specifically includes: Based on the completeness of data from each data source within the current evaluation period, the integrity score of each data source is determined using the main control processor. Based on the degree of time matching between each data source and the current evaluation period, the timeliness score of each data source is determined based on the main control processor; Based on whether each data source has a clear source path and verification path, the traceability score of each data source is determined based on the main control processor; Based on the degree of mutual verification between each data source and other data sources, the consistency score of each data source is determined based on the main control processor. Based on the integrity score, timeliness score, traceability score, and consistency score, and the corresponding sub-coefficients, the raw reliability value of each data source is calculated using the main control processor. R i ; Based on the original credibility values ​​of each data source R i The normalization process is performed based on the main control processor to make the sum of the weights of all data sources equal to 1, thus obtaining the data source weights of each data source. Original credibility value Ri The expression is: (2); In formula (2), α 1. α 2. α 3. α 4 represents the coefficients for each component, and the sum of these coefficients is 1. q i For the first i Integrity score of each data source t i For timeliness, l i For traceability score, s i For consistency score; The formula for calculating the data source weight is: (3); In formula (3), ρ i Indicates the first i The weight of each data source.

[0013] Furthermore, in step 6... u The fusion score of each evaluation indicator is calculated as follows: (4); In formula (4), Indicates the first u The combined score of the evaluation indicators For the first i The data source in the first u Standardized values ​​for each indicator ρ i Indicates the first i The weight of each data source, n Indicates the total number of data sources; Furthermore, step 7 specifically includes: Based on at least one of the enterprise's environmental governance maturity, system completeness, and policy adaptability, the main control processor generates environmental adjustment items; Based on at least one of the following criteria: the proportion of missing key indicators, the degree of cross-source conflict, or the number of process anomalies, a comprehensive penalty item is generated based on the main control processor. The comprehensive score of green innovation performance is calculated based on the main control processor, according to the weighted sum of the comprehensive weight of each evaluation indicator and the fusion score of each evaluation indicator, the environmental adjustment item and its adjustment coefficient, and the comprehensive penalty item and its penalty coefficient. The formula for calculating the comprehensive score of green innovation performance is as follows: (5); In formula (5), G This indicates the overall score for green innovation performance. m This represents the total number of evaluation indicators. This represents the comprehensive weight of the u-th evaluation indicator. Indicates the first u The combined score of the evaluation indicators E Indicates environmental adjustment items. η Indicates the environmental regulation coefficient. P This indicates a comprehensive penalty item. λ This represents the penalty coefficient.

[0014] Furthermore, step 8 specifically includes: The green innovation performance level is determined based on the overall score of the green innovation performance and the main control processor. Based on the fusion scores of various evaluation indicators, the main control processor generates sub-indicator analysis results and identifies the main weak indicators. Based on the data source weight, a reliable data source description is generated based on the main control processor. Based on the analysis results of sub-indicators and the main shortcomings, improvement suggestions are generated based on the main control processor.

[0015] Furthermore, step 9 specifically includes: The main control processor receives real green innovation results data for the next evaluation cycle. Real green innovation results data include green product transformation results, green patent implementation effects, resource consumption improvement effects, emission reduction completion status, and external governance results. Based on the deviation between the actual green innovation results data and the performance evaluation results generated in the current evaluation cycle, the main control processor determines whether the deviation exceeds a preset range, wherein the preset range is determined by the historical evaluation error mean, error standard deviation or preset tolerance threshold. When the deviation exceeds the preset range, the main control processor adjusts at least one of the following: the subjective-objective fusion coefficient, the credibility sub-coefficient of the data source, the adjustment coefficient of the environmental adjustment item, and the penalty coefficient of the comprehensive penalty item, so that the evaluation results of the next evaluation cycle are closer to the actual business performance.

[0016] Furthermore, this invention also proposes a green innovation performance evaluation system based on dynamic weight optimization, comprising: The system includes a data access module, a standardization processing module, an indicator construction module, a weight calculation module, a data credibility module, a comprehensive evaluation module, a result output module, a feedback correction module, and a main control processor. The data access module is used to receive multi-source data related to green innovation from internal enterprise systems, external databases and network platforms, and to complete source identification and preliminary classification; The standardization processing module is used to process data of different formats in a unified manner, including numerical normalization, text semantic representation, log statistical transformation, image feature extraction and time window alignment, so that data from different sources can enter the same evaluation framework. The indicator construction module is used to form a hierarchical structure of technical indicators, organizational indicators, environmental indicators and output indicators based on preset green innovation evaluation rules, and to select core indicators according to business scenarios. The weight calculation module is used to calculate the subjective weight, objective weight, and comprehensive weight of each evaluation indicator, and outputs the comprehensive weight to the comprehensive evaluation module. The data credibility module is used to score the integrity, timeliness, traceability and consistency of different data sources and generate data source weights; The comprehensive evaluation module is used to calculate the total score and sub-indicator scores of green innovation performance based on indicator weights, data source weights, indicator fusion scores, and environmental adjustment items. The results output module is used to display or output performance levels, sub-indicator evaluation conclusions, weak indicators, credible data explanations, and improvement suggestions; The feedback correction module is used to receive subsequent actual business results and correct relevant parameters based on the deviation results; The main control processor can be a server, industrial computer, edge computing device, or processor that supports floating-point operations, used to perform data scheduling, parameter calculation, process control, and result updating.

[0017] The beneficial effects of this invention are: 1) This invention changes the evaluation method that relies only on a few result indicators or a single manual score. It integrates the technological foundation, organizational input, environmental constraints and green output into the evaluation system, which can more comprehensively reflect the true state of green innovation of enterprises and avoid the one-sided evaluation that only looks at the results and not the process.

[0018] 2) This invention simplifies the dynamic weight optimization process. Without compromising technical sophistication, it significantly reduces the deployment burden caused by complex model stacking by analyzing the comprehensive weight of indicators, data source weight, indicator fusion score, comprehensive performance score, and feedback correction, making the system process faster.

[0019] 3) This invention introduces a data source credibility calculation mechanism, treating completeness, timeliness, traceability, and consistency as independent factors. This solves the problem of mixed use of different data sources and failure to analyze credibility differences in existing technologies, thereby improving the robustness and credibility of the evaluation results.

[0020] 4) This invention incorporates factors such as environmental ethics, green institutions, and policy embedding into the final performance results through environmental adjustment items, so that green innovation evaluation is no longer limited to technology and output, but can reflect the promoting effect of green governance on innovation performance, which is more in line with the actual law that green innovation activities are affected by environmental factors.

[0021] 5) This invention uses a main control processor to uniformly schedule the evaluation process, which not only facilitates the integrated operation of data collection, weight calculation, result output and parameter correction, but also facilitates deployment in existing enterprise information systems, meeting the actual needs of periodic measurement, process tracking and online indicator maintenance. Attached Figure Description

[0022] Figure 1 A flowchart of a green innovation performance evaluation method based on dynamic weight optimization; Figure 2 This is a structural block diagram of a green innovation performance evaluation system based on dynamic weight optimization. Figure 3 Optimize the flowchart for dynamic weights; Figure 4 Flowchart for calculating data source credibility; Figure 5 The flowchart was revised based on feedback. Detailed Implementation

[0023] like Figure 1 As shown, the steps of a green innovation performance evaluation method based on dynamic weight optimization described in this embodiment include: S1: Multi-source data access; Multi-source data includes, but is not limited to, operational data from enterprise business systems, project data from R&D management systems, cost and input data from financial systems, energy consumption and emission data from environmental governance systems, policy text information from policy databases, feedback data from social platforms, and supplementary data generated by park terminals, log systems, or image acquisition devices. Because these data differ in source, structure, time scale, and reliability, the main control processor needs to perform unified access and classification labeling before subsequent calculations. The classification labeling includes the data source, data type, collection time, and associated business objects to facilitate subsequent calculation of data source weights and result traceability.

[0024] S2: Data standardization and semantic mapping; This step is not limited to simple numerical scaling, but involves preprocessing based on data type. For numerical data, it primarily performs missing value imputation, outlier correction, and dimensional normalization. For textual data, it mainly performs semantic processing, keyword extraction, topic identification, and rule mapping. For log data, it primarily performs event aggregation, frequency statistics, and time window alignment. For image data, it primarily performs target feature extraction and structured representation. The processed data is then uniformly transformed into standardized values ​​suitable for evaluation calculations, forming a data matrix. This data matrix retains the richness of multi-source information while avoiding the incomparability issues caused by directly mixing different data formats.

[0025] S3: Indicator Construction and Core Indicator Determination; After the data matrix is ​​constructed, the system further enters the indicator construction phase. This invention establishes an evaluation indicator system from four levels: technical indicators, organizational indicators, environmental indicators, and output indicators. Technical indicators describe the enterprise's digital capabilities, infrastructure, collaborative platform capabilities, and green technology support level in green innovation. Organizational indicators describe the enterprise's investment in green innovation in terms of human resources, capital, systems, and management. Environmental indicators describe the policy support, environmental ethical constraints, green culture embedding, and social feedback received by the enterprise. Output indicators describe outcome indicators such as green patents, green products, green services, emission reduction effectiveness, and improved resource utilization. Through this multi-indicator structure, green innovation performance evaluation is no longer limited to the numerical results themselves, but can provide a more complete support for the process and external conditions.

[0026] S4: Calculation of indicator weights; like Figure 3 As shown, the dynamic weight optimization process in this invention mainly includes three stages: subjective weight generation, objective weight generation, and comprehensive weight fusion. First, subjective weights for each indicator are formed based on the evaluation objectives and industry experience. These subjective weights reflect the evaluators' or domain experts' judgments on the importance of different indicators, primarily used to reflect the differences in business priorities, industry preferences, and governance objectives in green innovation scenarios. These subjective weights can be generated using expert scoring methods, the analytic hierarchy process (AHP), or preset business rules. Next, objective weights for each indicator are calculated based on actual sample data. These objective weights reflect the information content, discriminative power, and relative contribution of each indicator at the data level, used to reduce subjective fluctuations caused by relying solely on experience-based weighting. These objective weights can be calculated based on the indicator's dispersion, information entropy, or variance contribution. Finally, the subjective and objective weights are synthesized according to a set fusion coefficient to obtain the final comprehensive weights for this round of evaluation.

[0027] In this invention, the first u The comprehensive weight of each evaluation indicator is expressed as follows: (1); In formula (1), w u Indicates the first u The final weight of each indicator, θ This represents the fusion coefficient between subjective and objective weights, with a value ranging from 0 to 1. For the first u The subjective weight of each indicator, For the first u The objective weights of each indicator. The meaning of this formula is that different business scenarios can be adjusted... θ The values ​​of these values ​​allow for different evaluation biases, prioritizing experience or data. When a company is in the early stages of green innovation exploration and has relatively limited historical data, the influence of subjective weights can be appropriately increased. Conversely, when a company has accumulated relatively rich operational and outcome data, the influence of objective weights can be appropriately increased. Through this simplified and clear weight generation method, this invention avoids the problem of excessive stacking of complex models while retaining the technical requirements for dynamic weight optimization.

[0028] S5: Calculating data source weights; like Figure 4 As shown, the data source credibility calculation process in this invention mainly includes two stages: generating the original credibility value and normalizing the data source weights. Since green innovation performance evaluation relies on multi-source data, and different data sources differ significantly in terms of completeness, timeliness, traceability, and consistency, a simple averaging approach cannot be used for synthesis on the same indicator. This invention first calculates the original credibility value for each data source, then normalizes the original credibility values ​​of each data source into data source weights, and finally uses these weights for the fusion of multi-source data under the same indicator.

[0029] No. i The original credibility values ​​of each data source are represented as follows: (2); In formula (2), α 1. α 2. α 3. α 4 represents the coefficients of each component, and their sum is 1. Let there be a total of... n The data source, the first i The integrity score of each data source is q i The timeliness score is t i Traceability score: l i Consistency score: s i .

[0030] The completeness score reflects whether there are any significant gaps in the data source within the current evaluation period. The timeliness score reflects the degree of time matching between the data source and the current evaluation task. The traceability score reflects whether the data has a clear source and verification path. The consistency score reflects whether the data can corroborate other data. This raw credibility value can directly quantify the differences in data quality from different sources, thus avoiding the equating of high-credibility data with low-credibility data.

[0031] After calculating the raw credibility values ​​for each data source, the system normalizes them to obtain the data source weights: (3); In formula (3), ρ i Indicates the first i The weight of each data source, n This represents the total number of data sources. This normalization step ensures that the sum of the weights of all data sources is 1, facilitating weighted fusion under the same metric.

[0032] S6: Calculation of integrated index score; For the u The system calculates the fusion score based on the weights of each data source and their corresponding standardized values, using one evaluation metric: (4); In formula (4), Indicates the first u The combined score of the evaluation indicators ρ i Indicates the first i The weight of each data source. i The data source in the first u The standardized value for each indicator z iu , n This indicates the total number of data sources.

[0033] S7: Calculation of the overall score for green innovation performance; After completion Figure 3 and Figure 4 After the weight calculation and indicator integration shown, the system enters the comprehensive performance calculation stage. Specifically, the main control processor calculates the comprehensive performance based on the combined weights of each indicator. w u Combined score of various indicators S u Calculate the overall score of green innovation performance GSimultaneously, this invention introduces environmental adjustment and comprehensive penalty items to reflect the corrective effects of factors such as green system construction, environmental ethics level, key data conflicts, or evaluation anomalies on the final result. The comprehensive performance score can be expressed as: (5); In formula (5), G This indicates the overall score for green innovation performance. m This represents the total number of evaluation indicators. w u Indicates the first u The overall weight of each evaluation indicator S u Indicates the first u The combined score of the evaluation indicators E Indicates environmental adjustment items. η Indicates the environmental regulation coefficient. P This indicates a comprehensive penalty item. λ This represents the penalty coefficient. The resulting comprehensive score can be directly mapped to a green innovation performance level, or it can be further broken down into technical indicator scores, organizational indicator scores, environmental indicator scores, and output indicator scores to identify a company's strengths and weaknesses in the green innovation chain. The environmental adjustment item... E The comprehensive penalty item is calculated by normalization of at least one of environmental governance maturity, institutional completeness, and policy adaptability. P It is determined based on at least one of the following: the proportion of missing key indicators, the degree of cross-source conflict, or the number of process anomalies.

[0034] S8: Evaluation results output and weakness diagnosis; S9: Feedback correction and parameter update; like Figure 5 As shown, this invention further incorporates a processor feedback correction process. This process aims to prevent the evaluation model from remaining static for extended periods, thereby improving its adaptability to changes in business operations and the environment. Specifically, after completing the green innovation performance evaluation for the current cycle, the system stores the evaluation results in historical records. When the actual performance results corresponding to the current evaluation object become available at a later time, the main control processor compares these actual performance results with the predictive evaluation results generated in the corresponding evaluation cycle and corrects the parameters used in the next evaluation cycle accordingly. If the deviation between the two exceeds a preset range, a parameter correction process is triggered. The preset range can be determined by the historical evaluation error mean, error standard deviation, or a preset tolerance threshold.

[0035] Feedback correction does not alter the main formula structure of this invention, but rather revolves around several key parameters, including the subjective-objective fusion coefficient, the data source credibility coefficient, the environmental adjustment coefficient, and the penalty coefficient. By correcting these parameters, the evaluation results of the next cycle can be made closer to actual business performance. For example, when the system finds a weak correlation between social feedback data and actual output for several consecutive cycles, the weight coefficient of this type of data in the credibility calculation can be appropriately lowered. When the system finds that the construction of corporate environmental systems has a more significant effect on subsequent performance improvement, the environmental adjustment coefficient can be appropriately increased. Through this lightweight correction method, this invention maintains the stability of the calculation structure while enabling the evaluation system to continuously learn and optimize. Specific Implementation Method Two like Figure 2 As shown in this embodiment, a green innovation performance evaluation system based on dynamic weight optimization includes: a data access module, a standardization processing module, an indicator construction module, a weight calculation module, a data credibility module, a comprehensive evaluation module, a result output module, a feedback correction module, and a main control processor. The modules communicate with each other according to predetermined data and control flows, and are uniformly scheduled by the main control processor. Specifically, the data access module is responsible for collecting multi-source data related to green innovation from internal enterprise systems, external databases, and network platforms. The standardization processing module is responsible for converting the collected data into standardized data in a unified format. The indicator construction module is responsible for establishing a green innovation performance evaluation indicator system and extracting a core indicator set according to preset rules. The weight calculation module is responsible for generating the comprehensive weight of each indicator. The data credibility module is responsible for calculating the credibility of different data sources and their normalized weights. The comprehensive evaluation module is responsible for calculating scores based on weights and indicator values. The result output module is responsible for generating evaluation conclusions, indicator analysis, and recommendation information. The feedback correction module is responsible for correcting relevant parameters based on subsequent actual results.

[0037] The core indicator set includes one or more of the following: the number of new energy facilities, the number of green patents granted, the coverage of smart scenarios, the proportion of R&D personnel, R&D investment intensity, corporate green governance maturity, the intensity of social green feedback, and the strength of green policy support. It also includes extended indicators to characterize management capabilities, technological capabilities, talent capabilities, and environmental ethics levels. These extended indicators are used to provide supplementary explanations for the green innovation performance evaluation results or to form environmental adjustment items. The core indicator set is selected from the candidate indicator set based on industry type, data availability, indicator discriminative power, and business scenario configuration rules.

[0038] The main control processor is located at the center of the entire system. This main control processor can be a server, industrial computer, edge computing device, or a processor that supports floating-point operations, and is used to perform data scheduling, parameter calculation, process control, and result updates. Through the unified control of the main control processor, the modules are no longer loosely connected and independent, but rather form a continuously operating and continuously correcting green innovation performance evaluation system.

[0039] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent substitutions, and improvements made to the above embodiments without departing from the scope of the present invention, based on the technical essence of the present invention and within the spirit and principles of the present invention, shall still fall within the protection scope of the present invention.

Claims

1. A green innovation performance evaluation method based on dynamic weight optimization, characterized in that, include: Step 1: Receive multi-source data related to green innovation activities based on the main control processor; Step 2: Based on the multi-source data, perform data preprocessing using the main control processor to form a unified data matrix; Step 3: Based on the data matrix and the main control processor, determine the technical indicators, organizational indicators, environmental indicators and output indicators, and construct a green innovation performance evaluation indicator system. The indicator system includes technical indicators, organizational indicators, environmental indicators and output indicators, and determine the core indicator set from it. Step 4: Based on the core indicator set, generate subjective and objective weights for each evaluation indicator using the main control processor, and then weight and fuse the subjective and objective weights to obtain the comprehensive weight of each evaluation indicator. Step 5: Based on the integrity, timeliness, traceability and consistency of each data source, calculate the integrity score, timeliness score, traceability score and consistency score respectively based on the main control processor, and perform weighted summation according to the preset sub-item coefficients to obtain the original credibility value of each data source. Normalize the original credibility value to obtain the data source weight of each data source. Step 6: Based on the standardized values ​​of each data source on each evaluation indicator in the core indicator set and the data source weights, calculate the fusion score of each evaluation indicator based on the main control processor; Step 7: Calculate the comprehensive score of green innovation performance based on the main control processor, according to the comprehensive weight of each evaluation indicator, the fusion score of each evaluation indicator, the environmental adjustment item, and the comprehensive penalty item. Step 8: Based on the comprehensive score of green innovation performance, output the green innovation performance evaluation results based on the main control processor, and generate sub-indicator analysis results, main shortcomings indicators, explanation of reliable data sources, and improvement suggestions; Step 9: Based on subsequent actual business feedback, adjust the subjective and objective fusion coefficient, the credibility sub-coefficient of the data source, the adjustment coefficient of the environmental adjustment item, and the penalty coefficient of the comprehensive penalty item based on the main control processor.

2. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, The multi-source data in step 1 includes enterprise business data, R&D data, financial data, environmental governance data, policy text data, social feedback data, images, and log data; Step 1 specifically includes: Based on the main control processor, it receives operational data from enterprise business systems, R&D activity data from R&D systems, input and cost data from financial systems, energy consumption and emission reduction data from environmental governance systems, policy text data from policy databases, feedback data from social platforms, and extracts structured or semi-structured data from park terminals, image acquisition equipment, or log systems. Based on the source, type, collection time, and associated business objects of the multi-source data, the main control processor classifies and identifies the multi-source data to form an original dataset with source traceability information.

3. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 2 specifically includes: Step 2.1: Based on the missing values, outliers, and dimensional differences of the numerical data in the original dataset, perform missing value imputation, outlier correction, and dimensional normalization using the main control processor; Step 2.2: Based on the semantic content of the text data in the original dataset, perform semantic processing, keyword extraction, and rule mapping using the main control processor; Step 2.3: Based on the event records of the log-type data in the original dataset, perform event aggregation, frequency statistics, and time window alignment using the main control processor; Step 2.4: Based on the visual information of the image data in the original dataset, perform target feature extraction and structured representation using the main control processor; Step 2.5: Based on the processing results of Steps 2.1-2.4, the main control processor maps all types of data into standardized values ​​to obtain a data matrix.

4. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 3 includes: Based on the data items in the data matrix that reflect the enterprise's green innovation technology foundation, technical indicators are determined based on the main control processor. These technical indicators are used to characterize the digital foundation, equipment foundation, and technological synergy capabilities upon which the enterprise's green innovation depends. Based on the data items reflecting the enterprise's green innovation investment in the data matrix, organizational indicators are determined based on the main control processor. These organizational indicators are used to characterize the enterprise's human resource investment, financial investment, institutional investment, and management coordination capabilities in green innovation. Based on the data items in the data matrix that reflect the external constraints and feedback of the enterprise, environmental indicators are determined based on the main control processor. These environmental indicators are used to characterize policy drivers, social feedback, environmental ethics, and external governance constraints. Based on the data items in the data matrix that reflect the results of enterprise green innovation, output indicators are determined based on the main control processor. These output indicators are used to characterize green patents, green products, green services, resource utilization improvements, and emission reduction results. Based on technical indicators, organizational indicators, environmental indicators, and output indicators, a core indicator set is formed by screening from the candidate indicator set using the main control processor. The core indicator set includes one or more of the following: number of new energy facilities, number of green patent authorizations, coverage of smart scenarios, proportion of R&D personnel, R&D investment intensity, maturity of corporate green governance, intensity of social green feedback, and strength of green policy support.

5. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 4 specifically includes: Based on the evaluation objectives and industry experience, subjective weights for each evaluation indicator are generated using one of the following methods: expert scoring, analytic hierarchy process, or preset business rules, based on the main control processor. The subjective weights reflect the experts' judgments on the relative importance of each indicator. Based on the sample data corresponding to the core indicator set, the main control processor calculates the objective weight of each evaluation indicator based on one of the indicator dispersion, information entropy, or variance contribution. The objective weight reflects the information difference and distinguishing ability of each indicator in the sample data. Based on the preset subjective-objective fusion coefficients, the subjective and objective weights are weighted and fused using the main control processor to generate the comprehensive weights for each evaluation indicator. w u ; u The overall weight of each indicator is: （1）； In formula (1), w u Indicates the first u The final weight of each indicator, θ This represents the fusion coefficient between subjective and objective weights, with a value ranging from 0 to 1. For the first u The subjective weight of each indicator, For the first u The objective weight of each indicator.

6. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 5 specifically includes: Based on the completeness of data from each data source within the current evaluation period, the integrity score of each data source is determined using the main control processor. Based on the degree of time matching between each data source and the current evaluation period, the timeliness score of each data source is determined based on the main control processor; Based on whether each data source has a clear source path and verification path, the traceability score of each data source is determined based on the main control processor; Based on the degree of mutual verification between each data source and other data sources, the consistency score of each data source is determined based on the main control processor. Based on the integrity score, timeliness score, traceability score, and consistency score, and the corresponding sub-coefficients, the original reliability value of each data source is calculated using the main control processor. R i ; Based on the original credibility values ​​of each data source R i The normalization process is performed based on the main control processor to make the sum of the weights of all data sources equal to 1, thus obtaining the data source weights of each data source. Original credibility value R i The expression is: （2）； In formula (2), α 1. α 2. α 3. α 4 represents the coefficients for each component, and the sum of these coefficients is 1. q i For the first i Integrity score of each data source t i For timeliness, l i For traceability score, s i For consistency score; The formula for calculating the data source weight is: （3）； In formula (3), ρ i Indicates the first i The weight of each data source, n This indicates the total number of data sources.

7. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 6 u The fusion score of each evaluation indicator is calculated as follows: （4）； In formula (4), Indicates the first u The combined score of the evaluation indicators For the first i The data source in the first u Standardized values ​​for each indicator ρ i Indicates the first i The weight of each data source, n Indicates the total number of data sources; Step 7 specifically includes: Based on at least one of the enterprise's environmental governance maturity, system completeness, and policy adaptability, the main control processor generates environmental adjustment items; Based on at least one of the following criteria: the proportion of missing key indicators, the degree of cross-source conflict, or the number of process anomalies, a comprehensive penalty item is generated based on the main control processor. The comprehensive score of green innovation performance is calculated based on the main control processor, according to the weighted sum of the comprehensive weight of each evaluation indicator and the fusion score of each evaluation indicator, the environmental adjustment item and its adjustment coefficient, and the comprehensive penalty item and its penalty coefficient. The formula for calculating the comprehensive score of green innovation performance is as follows: （5）； In formula (5), G This indicates the overall score for green innovation performance. m This represents the total number of evaluation indicators. Indicates the first u The overall weight of each evaluation indicator Indicates the first u The combined score of the evaluation indicators E Indicates environmental adjustment items. η Indicates the environmental regulation coefficient. P This indicates a comprehensive penalty item. λ This represents the penalty coefficient.

8. The green innovation performance evaluation method based on dynamic weight optimization according to claim 1, characterized in that, Step 8 specifically includes: The green innovation performance level is determined based on the overall score of the green innovation performance and the main control processor. Based on the fusion scores of various evaluation indicators, the main control processor generates sub-indicator analysis results and identifies the main weak indicators. Based on the data source weight, a reliable data source description is generated based on the main control processor. Based on the analysis results of sub-indicators and the main shortcomings, improvement suggestions are generated based on the main control processor.

9. The green innovation performance evaluation method based on dynamic weight optimization according to claim 8, characterized in that, Step 9 specifically includes: The main control processor receives real green innovation results data for the next evaluation cycle, including green product transformation results, green patent implementation effects, resource consumption improvement effects, emission reduction completion status, and external governance results. Based on the deviation between the actual green innovation results data and the performance evaluation results generated in the current evaluation cycle, the main control processor determines whether the deviation exceeds a preset range, wherein the preset range is determined by the historical evaluation error mean, error standard deviation or preset tolerance threshold. When the deviation exceeds the preset range, the main control processor adjusts at least one of the following: the subjective-objective fusion coefficient, the credibility sub-coefficient of the data source, the adjustment coefficient of the environmental adjustment item, and the penalty coefficient of the comprehensive penalty item, so that the evaluation results of the next evaluation cycle are closer to the actual business performance.

10. A green innovation performance evaluation system based on dynamic weight optimization, applied to the green innovation performance evaluation method based on dynamic weight optimization as described in any one of claims 1-9, comprising: The system includes a data access module, a standardization processing module, an indicator construction module, a weight calculation module, a data credibility module, a comprehensive evaluation module, a result output module, a feedback correction module, and a main control processor. The data access module is used to receive multi-source data related to green innovation from internal enterprise systems, external databases and network platforms, and to complete source identification and preliminary classification. The standardization processing module is used to uniformly process data of different formats, including numerical normalization, text semantic representation, log statistical transformation, image feature extraction and time window alignment, so that data from different sources can enter the same evaluation framework. The indicator construction module is used to form a hierarchical structure of technical indicators, organizational indicators, environmental indicators and output indicators based on preset green innovation evaluation rules, and to select core indicators according to business scenarios. The weight calculation module is used to calculate the subjective weight, objective weight and comprehensive weight of each evaluation indicator, and output the comprehensive weight to the comprehensive evaluation module. The data credibility module is used to score the integrity, timeliness, traceability and consistency of different data sources, and generate data source weights; The comprehensive evaluation module is used to calculate the total score and sub-indicator scores of green innovation performance based on indicator weights, data source weights, indicator fusion scores, and environmental adjustment items. The results output module is used to display or output performance level, sub-indicator evaluation conclusions, weak indicators, credible data explanations, and improvement suggestions; The feedback correction module is used to receive subsequent actual business results and correct relevant parameters based on the deviation results; The main control processor can be a server, industrial computer, edge computing device, or processor that supports floating-point operations, and is used to perform data scheduling, parameter calculation, process control, and result updating.

Images